Electromyography Medical Encyclopedia.pdf
Transcript of Electromyography Medical Encyclopedia.pdf
-
8/12/2019 Electromyography Medical Encyclopedia.pdf
1/12
See also GASTROINTESTINAL HEMORRHAGE; GRAPHIC RECORDERS.
ELECTROMAGNETIC FLOWMETER. SeeFLOWMETERS, ELECTROMAGNETIC.
ELECTROMYOGRAPHY
CARLO DE LUCA
Boston University
Boston, Massachusetts
INTRODUCTION
Electromyography is the discipline that deals with the
detection, analysis, and use of the electrical signal that
emanates from contracting muscles.This signal is referred to as the electromyographic
(EMG) signal, a term that was more appropriate in the
past than in the present. In days past, the only way to
capture the signal for subsequent study was to obtain a
graphic representation. Today, of course, it is possible to
store the signal on magnetic tape, disks, and electronics
De Luca, C.J. Electromyography. Encyclopedia ofMedical Devices and Instrumentation,(John G. Webster,
Ed.) John Wiley Publisher, 98-109, 2006.Copyright2006 John G. Webster.This material is used by permission of John Wiley & Sons, Inc."
-
8/12/2019 Electromyography Medical Encyclopedia.pdf
2/12
The novice in this field may well ask, why study elec-tromyography? Why bother understanding the EMG sig-nal? There are many and varied reasons for doing so. Evena superficial acquaintance with the scientific literature willuncover various current applications in fields such asneurophysiology, kinesiology, motor control, psychology,rehabilitation medicine, and biomedical engineering.
Although the state of the art provides a sound and richcomplement of applications, it is the potential of future
applications that generates genuine enthusiasm.
HISTORICAL PERSPECTIVE
Electromyography had its earliest roots in the custompracticed by the Greeks of using electric eels to shockailments out of the body. The origin of the shock thataccompanied this earliest detection and application ofthe EMG signal was not appreciated until 1666 when anItalian, Francesco Redi, realized that it originated frommuscle tissue (1). This relationship was later proved byLuigi Galvani (2) in 1791 who staunchly defended thenotion. During the ensuing six decades, a few investigatorsdabbled with this newly discovered phenomenon, but itremained for DuBois Reymond (3) in 1849 to prove that theEMG signal could be detected from human muscle during avoluntary contraction. This pivotal discovery remaineduntapped for eight decades awaiting the development oftechnological implements to exploit its prospects. This
interval brought forth new instruments such as the cath-ode ray tube, vacuum tube amplifiers, metal electrodes,and the revolutionary needle electrode which providedmeans for conveniently detecting the EMG signal. Thissimple implement introduced by Adrian and Bronk (4) in1929 fired the imagination of many clinical researchers
Guided by the work of Inman et al. (5), in the mid-1940sto the mid-1950s several investigations revealed a mono-tonic relationship between the amplitude of the EMG
signal and the force and velocity of a muscle contraction.This significant finding had a considerable impact: It dra-matically popularized the use of electromyographic studiesconcerned with muscle function, motor control, and kine-siology. Kinesiological investigations received yet anotherimpetus in the early 1960s with the introduction of wireelectrodes. The properties of the wire electrode were dili-gently exploited by Basmajian and his associates duringthe next two decades.
In the early 1960s, another dramatic evolution occurred
in the field: myoelectric control of externally poweredprostheses. During this period, engineers from severalcountries developed externally powered upper limb pros-theses that were made possible by the miniaturization ofelectronics components and the development of lighter,more compact batteries that could be carried by amputees.Noteworthy among the developments of externally pow-ered prostheses was the work of the Yugoslavian engineerTomovic and the Russian engineer Kobrinski, who in thelate 1950s and early 1960s provided the first examples of
such devices.In the following decade, a formal theoretical basis for
electromyography began to evolve. Up to this time, allknowledge in the field had evolved from empirical andoften anecdotal observations. De Luca (6,7) described amathematical model that explained many properties of thetime domainparameters of the EMGsignal, andLindstrom(8) described a mathematical model that explained manyproperties of the frequency domain parameters of the EMGsignal. With the introduction of analytical and simulation
techniques, new approaches to the processing of the EMGsignal surfaced. Of particular importance was the work ofGraupe and Cline (9), who employed the autoregressivemoving average technique for extracting information fromthe signal.
The late 1970s and early 1980s saw the use of sophis-ticated computer algorithms and communication theory todecompose the EMG signal into the individual electricalactivities of the muscle fibers (1012). Today, the decom-position approach promises to revolutionize clinical elec-tromyography and to provide a powerful tool forinvestigating the detailed control schemes used by thenervous system to produce muscle contractions. In thesame vein, the use of a thin tungsten wire electrode fordetecting the action potential from single fibers was popu-larized for clinical applications (13,14). Other techniquesusing the surface EMG signal, such as the use of medianand mean frequencies of the EMG signal to describe thefunctional state of a muscle and the use of the conduction
velocity of the EMG signal to provide information on themorphology of the muscle fibers began to take hold. For areview, see De Luca (15).
The 1990s saw the effective application of modern signalprocessing techniques for the analysis and use of the EMG
signal. Some examples are the use of time and frequency
ELECTROMYOGRAPHY 99
Figure 1. The EMG signal recorded with surface electrodes loca-
ted on the skin above the first dorsal interosseous muscle in the
hand. The signal increases in amplitude as the force produced by
the muscle increases.
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
3/12
systematic measurements of the muscle fiber conductionvelocity for measuring the severity of the Duchenne Dys-
trophy (17); the analysis of motor unit action potential delay
for locating the origin, the ending and the innervation zone
of musclefibers (18); and the application of timefrequencyanalysis of the EMG signal to thefield of laryngology (19).
New and exciting developments are on the horizon. For
example, the use of large-scale multichannel detection of
EMG signals for locating sources of muscle fiber abnorm-ality (20); application of neural networks to provide greater
degrees of freedom for the control of myoelectric prostheses
(21), and for the analysis of EMG sensors data for assessingthe motor activities and performance of sound subjects (22)
and Stroke patients (23). Yet another interesting develop-
ment is the emerging use of sophisticated Artificial Intelli-gence techniques for the decomposing the EMG signal (24).
The reader who is interested in more historical and factual
details is referred to the book Muscles Alive (25).
DESCRIPTION OF THE EMG SIGNAL
The EMG signal is the electrical manifestation of the
neuromuscular activation associated with a contracting
muscle. The signal represents the current generated by
the ionic flow across the membrane of the muscle fibersthat propagates through the intervening tissues to reach
the detection surface of an electrode located in the envir-
onment. It is a complicated signal that is affected by the
anatomical and physiological properties of muscles and the
control scheme of the nervous system, as well as the
characteristics of the instrumentation used to detect and
observe it. Some of the complexity is presented in Fig. 2
that depicts a schematic diagram of the main physiological,
anatomical and biochemical factors that affect the EMG
signal. The connecting lines in the diagram show the
interaction among three classes of factors that influence
influenced by one or more of the causative factors and inturn influence the deterministic factors that representphysical characteristics of the action potentials. For
further details see De Luca (26).
In order to understand theEMG signal, it is necessary to
appreciate some fundamental aspects of physiology. Mus-
cle fibers are innervated in groups called motor units,which when activated generate a motor unit action poten-
tial. The activation from the central nervous system is
repeated continuously for as long as the muscle is required
to generate force. This continued activation generates
motor unit action potential trains. These trains from theconcurrently active motor units superimpose to form the
EMG signal. As the excitation from the Central Nervous
System increases to generate greater force in the muscle, a
greater number of motor units are activated (or recruited)
and thefiring rates of all the active motor units increases.
Motor Unit Action Potential
The most fundamental functional unit of a muscle is called
the motor unit. It consists of an a-motoneuron and all the
muscle fibers that are innervated by the motoneuronsaxonal branches. The electrical signal that emanates from
the activation of the muscle fibers of a motor unit that arein the detectable vicinity of an electrode is called the motor
unit action potential (MUAP). This constitutes the funda-
mental unit of the EMG signal. A schematic representation
of the genesis of a MUAP is presented in Fig. 3. Note the
many factors that influence the shape of the MUAP. Someof these are (1) the relative geometrical relationship of the
detection surfaces of the electrode and the muscles fibers ofthe motor unit in its vicinity; (2) the relative position of the
detection surfaces to the innervation zone, that is, the
region where the nerve branches contact the musclefibers;(3) the size of the muscle fibers (because the amplitude ofthe individual action potential is proportional to the dia-
100 ELECTROMYOGRAPHY
CAUSATIVE
FACTORS
INTERMEDIATE DETERMINISTIC
EMG SIGNAL INTERPRETATION
EXTRINISIC
ELECTRODE:
-MOTOR POINT
-MUSCLE EDGE
-FIBER ORIENT.
-TENDON
INTRINSIC
NUMBER ACTIVE MU
FIBER DIAMETER
FIBER LOCATION
OTHER FACTORS
SUPERPOSITION
AMPLITUDE(RMS/ARV)
SPECTRALVARIABLES
(MEDIAN/MEAN FREQ.)
(FORCE-NETTORQUE)
MUSCLEACTIVATION
(ON/OFF)
MUSCLEFATIGUE
MUSCLE
MU FORCETWITCH
MU FIRINGRATE
NUMBERDETECTEDMU
MUAPDURATION
MUAPSHAPE
RECRUITMENTSTABILITY
MUSCLEFORCE
BIOCHEM.
-CONFIGURATION
FIBER TYPELACTIC ACID (pH)
ELECTRODE
SUBCUTANEOUSTISSUE
MU FIRING RATE(SYNCHRONIZATION)
SPATIALFILTERING
DETECTIONVOLUME
NUMBERACTIVEMU
CONDUCTIONVELOCITY
SIGNALCROSSTALK
MUAPAMPLITUDE
DIFF. ELECTRODE
MUSCLEFIBERINTERACTIONS
FILTER
BLOOD FLOW
.
.
.
.
.
.
.
.
.
.
.
.
.
...
.
.
.
.
.
.
..
.
.
.
...
.
.
.
.
....
.
.
.
.
..
.
CJ DeLuca
MU FORCE TWITCH
Figure 2. Relationship among the various factors that affect the EMG signal. [Reproduced with
Permission from C. J. De Luca, The Use of Surface Electromyography in Biomechanics. In the
Journal of Applied Biomechanics, Vol. 13(No 2): p 139, Fig. 1.]
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
4/12
The last two factors have particular importance in
clinical applications. Considerable work has been per-
formed to identify morphological modifications of the
MUAP shape resulting from modifications in the morpho-logy of the musclefibers (e.g., hypertrophy and atrophy) orthe motor unit (e.g., loss of muscle fibers and regenerationof axons). Although usage of MUAP shape analysis is
common practice among neurologists, interpretation of
the results is not always straightforward and relies heavily
on the experience and disposition of the observer.
Motor Unit Action Potential Train
The electrical manifestation of a MUAP is accompanied by
a contractile twitch of the muscle fibers. To sustain amuscle contraction, the motor units must be activated
repeatedly. The resulting sequence of MUAPs is called a
motor unit action potential train (MUAPT). The waveform
of the MUAPs within a MUAPT will remain constant if the
geometric relationship between the electrode and theactive muscle fibers remains constant, if the propertiesof the recording electrode do not change, and if there are
no significant biochemical changes in the muscle tissue.Biochemical changes within the muscle can affect the
conduction velocity of the muscle fiber and the filteringproperties of the muscle tissue.
The MUAPT may be completely described by its inter-
pulse intervals (the time between adjacent MUAPs) and
the waveform of the MUAP. Mathematically, the inter-
pulse intervals may be expressed as a sequence of Diracdelta impulsesdit convoluted with afilterht that repre-sents the shape of the MUAP. Figure 4 presents a graphic
representation of a model for the MUAPT. It follows that
the MUAPT, uit can be expressed as
where
tk Xkl1
xl for k; l 1; 2; 3; . . . ;n
In the above expression,tkrepresents the time locations of
the MUAPs,x represents the interpulse intervals, n is the
total number of interpulse intervals in a MUAPT, and i,k,
and l are integers that denote specific events.
By representing the interpulse intervals as a renewalprocess and restricting the MUAP shape so that it is
invariant throughout the train, it is possible to derive
the approximations
Mean rectified value
Efjuit;Fjgffilit;F
Z 10
jhitjdt
Mean squared value
MSfjuit;Fjgffi lit;FZ
1
0h2i tdt
where F is the force generated by the muscle and is the
firing rate of the motor unit.
The power density spectrum of a MUAPT was derived
from the above formulation by LeFever and De Luca [(27)
and independently by Lago and Jones (28)]. It can be
expressed as
Sui v; t;F Sdi v; t;FjHijvj2lit;Ff1jMjv;t;Fj
2g
12 RealfMjv;t;FgjMjv;t;Fj2fjHijvj
2g
for 6 0where is the frequency in radians per second,Hi jvis
ELECTROMYOGRAPHY 101
Detectionsite
MotoneuronMuscle fiber
Geometricalarrangementof electrodes
andactivefibers
Tissueand
electrodefilter
functions
Superpositionof
actionpotentials
h(t)
1
2
i
n
=
+
+
+
Figure 3. Schematic representation of the generation of the
motor unit action potential.
FOURIER TRANSFORM
FREQUENCY FREQUENCY FREQUENCY
=
tk
Xt
h(t)u(t)
t
d(t tk)
P
Figure 4. Model for a motor unit action potential train (MUAPT)
and the corresponding Fourier transform of the interpulse inter-
vals (IPIs), the motor unit actions potentials (MUAP), and the
MUAPT.
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
5/12
The EMG Signal
The EMG signal may be synthesized by linearly summing
the MUAPTs. This approach is expressed in the equation
mt;F Xpi1
uit;F
and is displayed in Fig. 5, where 25 mathematically gen-
erated MUAPTs were added to yield the signal at the
bottom. This composite signal bears striking similarity
to the real EMG signal.
From this concept, it is possible to derive expressions for
commonly used parameters: mean rectified value, root-mean-squared (rms) value, and variance of the rectified
EMG signal. The interested reader is referred to MusclesAlive(25).
Continuing with the evolution of the model, it is possible
to derive an expression for the power density spectrum of
the EMG signal:
Smv; t;F Rv;dXpFi1
Sui v; t XqFi; j1
i6j
Sui ujv; t
264
375
whereRv
;d
Ksin2
vd
=2y
is the bipolar electrode fil-
ter function; d is the distance between detection surfaces of
the electrode; is the angular frequency;y is the conduction
velocity along the muscle fibers; Sui v is the power densityof the MUAPT, uit; Suiuj v is the cross-power densityspectrum MUAPTsuitandujt;pis the total number ofMUAPTs that constitute the signal; and qis the number of
MUAPTs with correlated discharges.
Lindstrom (8), using a dipole model, arrived at another
expression for the power density spectrum:
Smv; t;F Rv;d 1y2t;FG vd2yt;F
h iThis representation explicitly denotes the interconnection
between the spectrum of the EMG signal and the conduc-
tion velocity of the muscle fibers. Such a relationship is
ofh(t) as seen by the two detection surfaces of a stationary
bipolar electrode.
ELECTRODES
Two main types of electrodes are used to detect the EMG
signal: one is the surface (or skin) electrode and the other is
the inserted (wire or needle) electrode. Electrodes are
typically used singularly or in pairs. These configurationsare referred to as monopolar and bipolar, respectively.
Surface Electrodes
There are two categories of surface electrode: passive andactive. Passive electrode consists of conductive (usually
metal) detection surface that senses the current on the
skin through its skin electrode interface. Active electrodes
contain a high input impedance electronics amplifierin the
same housing as the detection surfaces. This arrangement
renders it less sensitive to the impedance (and therefore
quality) of the electrodeskin interface. The current trendis towards active electrodes.
The simplest form of passive electrode consists of silver
disks that adhere to the skin. Electrical contact is greatlyimproved by introducing a conductive gel or paste between
the electrode and skin. The impedance can be further
reduced by removing the dead surface layer of the skin
along with its protective oils; this is best done by light
abrasion of the skin.
The lack of chemical equilibrium at the metal electrolyte
junction sets up a polarization potential that may vary with
temperaturefluctuations, sweat accumulation, changes inelectrolyte concentration of the paste or gel, relative move-
ment of the metal and skin, as well as the amount ofcurrentflowing into the electrode. It is important to notethat the polarization potential has both a direct current
(dc) and an alternating current (ac) component. The ac
component is greatly reduced by providing a reversible
chloride exchange interface with the metal of the electrode.
Such an arrangement is found in the silversilver chlorideelectrodes. This type of electrode has become highly pop-
ular in electromyography because of its light mass (0.25 g),
small size (
-
8/12/2019 Electromyography Medical Encyclopedia.pdf
6/12
(typically, 3 or 4 pF). The advent of modern microelectro-
nics has made possible the construction of amplifiershoused in integrated circuitry which have the required
input impedance and associated necessary characteristics.
An example of such an electrode is presented in Fig. 6. Thisgenre of electrodes was conceptualized and first con-structed at the NeuroMuscular Research Laboratory at
Childrens Hospital Medical Center, Boston, MA in the
late 1970s. They each have two detection surfaces and
associated electronic circuitry within their housing.
The chief disadvantages of surface electrodes are that
they can be used effectively only with superficial musclesand that they cannot be used to detect signals selectively
from small muscles. In the latter case, the detection of
cross-talksignals from other adjacent muscles becomes aconcern. These limitations are often outweighed by their
advantages in the following circumstances:
1. When representation of the EMG signal correspond-
ing to a substantial part of the muscle is required.
2. In motor behavior studies, when the time of activa-
tion and the magnitude of the signal contain the
required information.
3. In psychophysiological studies of general grossrelaxation of tenseness, such as in biofeedback
research and therapy.
4. In the detection of EMG signals for the purpose of
controlling external devices such as myoelectrically
controlled prostheses and other like aids for the
physically disabled population.
5. In clinical environments, where a relatively simple
assessment of the muscleinvolvement is required,for
example, in physical therapy evaluations and sports
medicine evaluations.6. Where the simultaneous activity or interplayof activ-
ity is being studied in a fairly large group of muscles
under conditions where palpation is impractical, for
example, in the muscles of the lower limb during
Needle Electrodes
By far, the most common indwelling electrode is the needle
electrode. A wide variety is commercially available. (see
Fig. 7). The most common needle electrode is the con-centric electrode used by clinicians. This monopolar con-figuration contains one insulated wire in the cannula. The
tip of the wire is bare and acts as a detection surface. Thebipolar configuration contains a second wire in the cannulaand provides a second detection surface. The needle elec-
trode has two main advantages. One is that its relatively
small pickup area enables the electrode to detectindividual
MUAPs during relatively low force contractions. The other
is that the electrodes may be conveniently repositioned
within the muscle (after insertion) so that new tissue
territories may be explored or the signal quality may be
improved. These amenities have naturally led to the devel-
opment of various specialized versions such as the multi-
filar electrode developed by Buchthal et al. (29), the planarquadrifilar electrode of De Luca and Forrest (30), the singlefiber electrode of Ekstedt and Stalberg (13), and the macro-electrode of Stalberg (14). The single-fiber electrode con-
ELECTROMYOGRAPHY 103
Figure 6. Examples of active surface electrode in bipolar config-
urations from Delsys Inc. The spacing between the bars is 10 mm,
the length of the bars is 10 mm and the thickness is 1 mm. Theseelectrodes do not require any skin preparation or conductive paste
or gels.
a
b
c
d
e
f
g
Figure 7. Examples of various needle electrodes: (a) A solid tip
single-fiber electrode.If itis sufficientlythin,it canbe inserted into
a nerve bundle and detect neuroelectrical signals. (b) Concentric
needle with one monopolardetectionsurface formedby thebeveled
cross-section of centrally located wire typically 200 mm in diame-
ter. Commonly used in clinical practice. (c) Bipolar needle elec-
trode with two wires exposed in cross-section, typically 100 mm indiameter. Used in clinical practice. (d) Single-fiber electrode with
25 mm diameter wire. Used to detect the activity of individual
muscle fibers. (e) Macroelectrode with 25 mm diameter wire and
with the cannula of the needle used as a detection surface. Used to
detect the motor unit action potential from a large portion of the
motor unit territory. (f) Quadrifilar planar electrode with four
50 mm wires located on the corners of a square 150mm apart
(center to center). Used for multiple channel recordings and in
EMG signal decomposition technique. (g) Multifilar electrode con-
sisting of a row of wires, generally used to study the motor unit
territory.
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
7/12
proven to be useful for neurological examinations of dein-
nervated muscles. Examples of these electrodes may be
seen in Fig. 7.
Wire Electrodes
Since the early 1960s, this type of electrode has been
popularized by Basmajian and Stecko (31). Similar elec-
trodes that differ only in minor details of construction were
developed independently at about the same time by other
researchers. Wire electrodes have proved a boon to kine-
siological studies because they are extremely fine, they areeasily implanted and withdrawn from skeletal muscles,
and they are generally less painful than needle electrodes
whose cannula remains inserted in the muscle throughoutthe duration of the test.
Wire electrodes may be made from any small diameter,
highly nonoxidizing, stiff wire with insulation. Alloys of
platinum, silver, nickel, and chromium are typically used.
Insulations, such as nylon, polyurethane, and Teflon, areconveniently available. The preferable alloy is 90% plati-
num, 10% iridium; it offers the appropriate combination of
chemical inertness, mechanical strength, stiffness and
economy. The Teflon and nylon insulations are preferredbecause they add some mechanical rigidity to the wires,
making them easier to handle. The electrode is constructed
by inserting two insulated fine (25100 mm in diameter)wires through the cannula of a hypodermic needle.
Approximately 12 mm of the distal tips of the wire isdeinsulated and bent to form two staggered hooks (see
Fig. 8for completed version). The electrode is introduced
into the muscle by inserting the hypodermic needle and
then withdrawing it. The wires remain lodged in the
muscle tissues. They may be removed by gently pullingthem out: They are so pliable that the hooks straighten out
on retraction.
In kinesiological studies, where the main purpose of
using wire electrodes is to record a signal that is propor-
tional to the contraction level of muscle, repositioning of
the electrode is not important. But for other applications,
such as recording distinguishable MUAPTs, this limitation
is counterproductive. Some have used the phrase poke andhopeto describe the standard wire electrode technique for
this particular application. Another limitation of the wireelectrode is its tendency to migrate after it has been
inserted, especially during the first few contractions ofthe muscle. The migration usually stops after a few con-
tractions. Consequently, it is recommended to perform a
half dozen or so short duration contraction before the
actual recording session begins.
Electrode Maintenance
Proper usage of wire and needle electrodes requires con-
stant surveillance of the physical and electrical character-
istics of the electrode detection surfaces. Particularattention should be given to keeping the tips free of debris
and oxidation. The reader is referred to the book Muscles
Alive(25)for details on these procedures as well as sugges-
tions for sterilization.
How to Choose the Proper Electrode
The specific type of electrode chosen to detect the EMGsignal depends on the particular application and the con-
venience of use. The application refers to the informationthat is expected to be obtained from the signal; for example,
obtaining individual MUAPs or the gross EMG signal
reflecting the activity of many muscle fibers. The conve-nience aspect refers to the time and effort the investigator
wishes to devote to the disposition of the subject or patient.
Children, for example, are generally resistant to having
needles inserted in their muscles.
The following electrode usage is recommended. The
reader, however, should keep in the mind that crossover
applications are always possible for specific circumstances.
Surface Electrodes
Time force relationship of EMG signals.
Kinesiological studies of surface muscles.
Neurophysiological studies of surface muscles.
Psychophysiological studies.
Interfacing an individual with external electromechan-
ical devices.
Needle Electrode
MUAP characteristics.
Control properties of motor units (firing rate, recruit-ment, etc.).
Exploratory clinical electromyography.
Wire Electrodes
Kinesiological studies of deep muscles.Neurophysiological studies of deep muscles.
Limited studies of motor unit properties.
Comfortable recording procedure from deep muscles.
Where to Locate the Electrode
The location of the electrode should be determined by three
important considerations: (1) signal/noise ratio, (2) signal
stability (reliability), and (3) cross-talk from adjacent mus-
cles. The stability consideration addresses the issue of the
modulation of the signal amplitude due to relative move-
ment of the active fibers with respect to the detectionsurfaces of the electrode. The issue of cross-talk concerns
the detection by the electrode of signals emanating from
104 ELECTROMYOGRAPHY
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
8/12
is so selective that it detects only signals from nearby
musclefibers. Because the musclefibers of different motorunits are scattered in a semirandom fashion throughout
the muscle, the location of the electrode becomes irrelevant
from the point of view of signal quality and information
content. The stability of the signal will not necessarily be
improved in any one location. Nonetheless, it is wise to
steer clear of the innervation zone so as to reduce the
probability of irritating a nerve ending.
All the considerations that have been discussed for
needle electrodes also apply to wire electrodes. In this case,
any complication will be unforgiving in that the electrode
may not be relocated. Since the wire electrodes have a
larger pickup area, a concern arises with respect to how thelocation of the insertion affects the stability of the signal.
This question is even more dramatic in the case of surface
electrodes.
For surface electrodes, the issue of cross-talk must be
considered. Obviously, it is not wise to optimize the signal
detected, only to have the detected signal unacceptably con-
taminated by an unwanted source. A second consideration
concerns the susceptibility of the signal to the architecture of
the muscle. Both the innervation zone and the tendon muscle
tissue interface have been found to alter the characteristics ofthe signal. It is suggested that the preferred location of an
electrode is in the region halfway between the center of the
innervation zone and the further tendon.See the review article
by De Luca (12) for additional details.
SIGNAL DETECTION: PRACTICAL CONSIDERATIONS
When attempting to collect an EMG signal, both the novice
and the expert should remember that the characteristics of
the observed EMG signal are a function of the apparatus
used to acquire the signal as well as the electrical current
that is generated by the membrane of the muscle fibers.The distortion of the signal as it progresses from thesource to the electrode may be viewed as a filteringsequence. An overview of the major filtering effects ispresented in Fig. 9. A brief summary of the pertinent facts
follows. The reader interested in additional details is
referred to Muscles Alive (25).
Electrode Configuration
The electrical activity inside a muscle or on the surface of
the skin outside a muscle may be easily acquired by placing
an electrode with only one detection surface in eitherenvironment and detecting the electrical potential at this
point with respect to a reference electrode located in anenvironment that either is electrically quiet or contains
electrical signals unrelated to those being detected. (Unre-latedmeans that the two signals have minimal physiolo-gical and anatomical associations.) A surface electrode is
commonly used as the reference electrode. Such an
arrangement is called monopolar and is at times used in
clinical environments because of its relative technical sim-
plicity. A schematic arrangement of the monopolar detec-tion configuration may be seen in Fig. 10. The monopolarconfiguration has the drawback that it will detect all theelectrical signals in the vicinity of the detection surface;
this includes unwanted signals from sources other than the
muscle of interest.
The bipolar detection configuration overcomes this lim-itation (see Fig. 10). In this case, two surfaces are used to
detect two potentials in the muscle tissue of interest each
with respect to the reference electrode. The two signals are
then fed to a differential amplifier which amplifies the
ELECTROMYOGRAPHY 105
Tissue(s)(low pass filter)
(anisotropy)
Electrode-electrolyteinterface
(high pass filter)
Bipolarelectrode
configuration(bandpass filter)
Amplifier(bandpass filter)Recorder(bandpass filter)ObservableEMG signal
PhysiologicalEMG signal
Figure 9. Block diagram of all the major aspects of the signal
acquisition procedure. Note the variety of physical properties that
act as filters to the EMG signal before it can be observed The term
Amp.
Detection
electrode
Detectionelectrode
m+ n
m1+ n m2+ n
Muscle
Muscle
Electricallyunrelatedtissue
Electricallyunrelated
tissue
Reference
electrode
Referenceelectrode
EMG Sig.
EMG Sig.Diff.amp.
+
(a)
(b)
Figure 10 (a) Monopolar detection arrangement (b) Bipolar
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
9/12
difference of the two signals, thus eliminating any com-mon mode components in the two signals. Signals ema-nating from the muscle tissue of interest near the detection
surface will be dissimilar at each detection surface because
of the localized electrochemical events occurring in the
contracting muscle fibers, whereas ac noise signals ori-ginating from a more distant source (e.g., 50 or 60 Hz
electromagnetic signals radiating from power cords, out-
lets, and electrical devices) and dc noise signals (e.g.,polarization potentials in the metal electrolyte junction)
will be detected with an essentially similar amplitude at
both detection surfaces. Therefore, they will be subtracted,
but not necessarily nullified prior to being amplified. The
measure bf the ability of the differential amplifier to elim-inate the common mode signal is called the common moderejection ratio.
Spatial Filtering
1. As the signal propagates through the tissues, the
amplitude decreases as a function of distance. The
amplitude of the EMG signal decreases to approxi-
mately 25% within 100 mm. Thus, an indwelling
electrode will detect only signals from nearby muscle
fibers.
2. Thefiltering characteristic of the muscle tissues is afunction of the distance between the active muscle
fibers and the detection surface(s) of the electrode. Inthe case of surface electrodes, the thickness of the
fatty and skin tissues must also be considered. The
tissues behaves as a low passfilter whose bandwidthand gain decrease as the distance increases.
3. The muscle tissue is anisotropic. Therefore, the orien-
tation of the detection surfaces of the electrode withrespect to the length of the muscle fibers is critical.
Electrode Electrolyte Interface
1. The contact layer between the metallic detection
surface of the electrode and the conductive tissue
forms an electrochemical junction that behaves as a
high pass filter.
2. The gain and bandwidth will be a function of the areaof the detection surfaces and any chemical electrical
alteration of the junction.
Bipolar Electrode Configuration
1. This configuration ideally behaves as a bandpassfilter; however, this is true only if the inputs to theamplifier are balanced and the filtering aspects of theelectrode electrolyte junctions are equivalent.
2. A larger interdetection surface spacing will render alower bandwidth. This aspect is particularly signifi-cant for surface electrodes.
3. The greater the interdetection surface spacing, the
greater the susceptibility of the electrode to detecting
4. An interdetection surface spacing of 1.0 cm is recom-
mended for surface electrodes.
Amplifier Characteristics
1. These should be designed and/or set for values that
will minimally distort the EMG signal detected by
the electrodes.
2. The leads to the input of the amplifier (actually, thefirst stage of the amplification) should be as short aspossible and should not be susceptible to movement.
This may be accomplished by building the first stageof the amplifier (the preamplifier) in a small config-uration which should be located near (within 10 cm)
the electrode. For surface EMG amplifiers thefirst stage is often located in the housing of theelectrodes.
3. The following are typical specifications that can beattained by modern dayelectronics. It is worth noting
that the values below will improve as more advanced
electronics components become available in the
future.
(a) Common-mode input impedance: As large as possible
(typically> 1015 V in parallel with < 7 pF).
(b) Common mode rejection ratio: > 85 dB.
(c) Input bias current: as low as possible (typically< 5 fA).
(d) Noise (shorted inputs)< 1.5 mV rms for 20500 Hzbandwidth.
(e) Bandwidth in hertz (3 dB points for 12 dB/octave or
more rolloff):
Surface electrodes 20500Wire electrodes 202,000
Monopolar and bipolar
needle electrodes for general use
205,000
Needle electrodes for signal
decomposition
1,00010,000
Singlefiber electrode 1,00010,000
Macroelectrode 205,000
An example of an eight-channel modern surface EMG
amplifier is presented in Fig. 11. Such systems are
106 ELECTROMYOGRAPHY
Figure 11. An eight-channel surface EMG system from Delsys
Inc The dimensions of this device (205 108 57 mm) are typical
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
10/12
availablein configurations of various channels up to 32,but8 and 16 channel versions are most common.
Recording Characteristics
The effective or actualbandwidth of the device or algorithm
that is used to record or store the signal must be greater
than that of the amplifiers.
Other Considerations
1. It is preferable to have the subject, the electrode, and
the recording equipment in an electromagnetically
quiet environment. If all the procedures and cautions
discussed in this article are followed and heeded,
high quality recordings will be obtained in the elec-
tromagnetic environments found in most institu-
tions, including hospitals.
2. In the use of indwelling electrodes, great caution
should be taken to minimize (eliminate, if possible)
any relative movement between the detection sur-
faces of the electrodes and the muscle fibers. Relativemovements of 0.1 mm may dramatically alter the
characteristics of the detected EMG signal and may
possibly cause the electrode to detect a differentmotor unit population.
SIGNAL ANALYSIS TECHNIQUES
The EMG signal is a time and force (and possibly other
parameters) dependent signal whose amplitude varies in a
random nature above and below the zero value. Thus,
simple average aging of the signal will not provide any
useful information.
Rectification
A simple method that is commonly used to overcome the
above restriction is to rectify the signal before performing
mode pertinent analysis. The process of rectificationinvolves the concept of rendering only positive deflectionsof the signal. This may be accomplished either by eliminat-
ing the negative values (half-wave rectification) or byinverting the negative values (full-wave rectification).
The latter is the preferred procedure because it retainsall the energy of the signal.
Averages or Means of Rectified Signals
The equivalent operation to smoothing in a digital sense is
averaging. By taking the average of randomly varying
values of a signal, the larger fluctuations are removed,thus achieving the same results as the analog smoothing
operation. The mathematical expression for the average or
mean of the rectified EMG signal is
jmtjtj ti 1tj ti
Z tjti
jmtjdt
where ti and tj are the points in time over which the
The preceding expression will provide only one value
over the time windowTtj ti. To obtain the time varyingaverage of a complete record of a signal, it is necessary to
move the time window Tduration along the record. This
operation is referred to as moving average.
jmtj 1T
Z tTt
jmtjdt
Like the equivalent operation in the analogue sense, this
operation introduces a lag; that is, Ttime must pass before
the value of the average of the T time interval can be
obtained. In most cases, this outcome does not present a
serious restriction, especially if the value of T is chosen
wisely. For typical applications, values ranging from 100 to200 ms are suggested. It should be noted that shorter time
windows, T, yield less smooth time dependent average
(mean) of the rectified signal.
Integration
The most commonly used and abused data reduction pro-
cedure in electromyography is integration. The literature
of the past three decades is swamped with improper usage
of this term, although happily within the past decade it ispossible tofind increasing numbers of proper usage. Whenapplied to a procedure for processing a signal, the temp
integration has a well-defined meaning that is expressed ina mathematical sense. It applies to a calculation that
obtains the area under a signal or a curve. The units of
this parameter are volt seconds (Vs). It is apparent that anobserved EMG signal with an average value of zero will
also have a total area (integrated value) of zero. Therefore,
the concept of integration may be applied only to the
rectified value of the EMG signal.
Ifjmtjg
Z tTt
jmtjdt
Note that the operation is a subset of the procedure of
obtaining the average rectified value. Since the rectifiedvalue is always positive, the integrated rectified value willincrease continuously as a function of time. The only
difference between the integrated rectified value andthe average rectified value is that in the latter case thevalue is divided by T, the time over which the average is
calculated. If a sufficiently long integration time T ischosen, the integrated rectified value will provide asmoothly varying measure of the signal as a function of
time. There is no additional information in the integrated
rectified value.
Root-Mean-Square (rms) Value
Mathematical derivations of the time and force dependent
parameters indicate that the rms value provides more a
more rigorous measure of the information content of the
signal because it measures the energy of the signal. Its use
in electromyography, however, has been sparse in the past.
The recent increase is due possibly to the availability of
ELECTROMYOGRAPHY 107
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
11/12
operations described by the term in reverse order; that is,
rms fmtg 1TZ tT
t
m2tdt !1=2
This parameter is recommended above the others.
Zero Crossings and Turns Counting
This method consists of counting the number of times per
unit time that the amplitude of the signal contains either a
peak or crosses a zero value of the signal. It was popular-
ized in electromyography by Williston (32). The relative
ease with which these measurements could be obtained
quickly made this technique popular among clinicians.Extensive clinical applications have been reported, some
indicating that discrimination may be made between myo-
pathic and normal muscle; however, such distinctions are
usually drawn on a statistical basis.
This technique is not recommended for measuring the
behavior of the signal as a function of force (when recruit-
ment or derecruitment of motor units occurs) or as a
function of time during a sustained contraction. Lindstrom
et al. (33)showed that the relationship between the turns
or zeros and the number of MUAPTs is linear for low levelcontractions. But as the contraction level increases, the
additionally recruited motor units contribute MUAPTs to
the EMG signal. When the signal amplitude attains the
character of Gaussian random noise, the linear proportion-
ality no longer holds.
Frequency Domain Analysis
Analysis of the EMG signal in the frequency domain
involves measurements and parameters that describe spe-cific aspects of the frequency spectrum of the signal. FastFourier transform techniques are commonly available and
are convenient for obtaining the power density spectrum of
the signal.
Three parameters of the power density spectrum maybe
conveniently used to provide useful measures of the spec-
trum. They are the median frequency, the mean frequency,
and the bandwidth of the spectrum. Other parameters,
such as the mode frequency and ratios of segments of the
power density spectrum, have been used by some investi-gators, but are not considered reliable measures given the
inevitably noisy nature of the spectrum. The median
frequency and the mean frequency are defined by theequations: Z fmed
0
Smfdf
Z 1fmed
Smfdf
fmean Z f
0
fSmfdfZ f
0
Smfdf
where Smf is the power density spectrum of the EMGsignal. Stulen and De Luca (34)performed a mathematical
analysis to investigate the restrictions in estimating var-
ious parameters of the power density spectrum. The med-
larly useful when a signal is obtained during low level
contractions where the signal to-noise ratio may be < 6.The above discussion on frequency spectrum parameters
removes temporal information from the calculated para-
meters. This approach is appropriate for analyzing signals
that are stationary or nearly stationary, such as those
emanating from isometric, constant-force contractions.
Measurement of frequency parameters during dynamic con-
tractions requirestechniques thatretainthe temporalinfor-
mation. During the past decade timefrequency analysestechniques haveevolvedin the field of Electromyography, asthey have in the realm of other biosignals such as ECG and
EEG. Early among the researchers to apply these techni-
ques to the EMG signal were Contable et al. (35) whoinvestigated the change in the frequency content of EMG
signals during high jumps, and Roark et al. (19) who inves-
tigated the movement of the thyroarytenoid muscles during
vocalization. In both these applications, the timefrequency
techniques were essential because they investigated mus-
cles that contracted dynamically and briefly.Much of the work presented here is adapted, with
permission, from Refs. 25, pp. 38, 58, 68, 74, and 81. The
author thanks Williams & Wilkens for permission to
extract this material.
BIBLIOGRAPHY
Cited References
1. Biederman W. Electrophysiology. 1898.
2. Galvani L. De Viribus Electricitatis. (R. Green, Transl.)
London and New York: Cambridge University Press; 1953.
3. Du Bois RE. Untersuchungen uber theirische electricitat. 2,
2nd P. Berlin: Verlag von G. Reimer; 1849.4. Adrian ED, Bronk DW. J Physiol (London) 1929;67:19.
5. Inman VT, Sauders JBCM, Abbott LC J. Bone Jt Surg
1944;26:1.
6. De LucaCJ. MS [dissertation].University of NewBrunswick;
1968.
7. De Luca CJ. Biol Cybernet 1975;19:159.
8. Lindstrom LR. On the Frequency Spectrum of EMG Signals.
TechnicalReport,ResearchLaboratory of Medical Electronics.
Gothenburg, Sweden: Chalmers University of Technology;
1970.
9. Graupe D, Cline WK. IEEE Trans Syst Man Cybernet SMC1975;5:252.
10. LeFever RS, De Luca CJ. Proceedings of the 8th Annual
Meeting of Social Neuroscience; 1985. p 299.
11. LeFever RS, De Luca CJ. IEEE Trans Biomed Eng BME
1982;29:149.
12. McGill KC, Cummins KL, Dorfman LJ. IEEE Trans Biomed
Eng 1985;32:470477.
13. Ekstedt J, Stalberg E. In: Desmedt JE, editor. New Develop-
ment EMGClinical Neurophysiology1. S.Karger; 1973.p 84.
14. Stalberg EJ. Neurol Neurosurg Psychiat 1980;43:475.
15. De Luca CJ. CRC Crit Rev Biomed Eng 1984;11:251279.16. Roy SH, De LucaCJ, EmleyMC.J Rehab Res Dev 1997;34(4):
405414.
17. Knaflitz M, Balestra G, Angelini C, Cadaldini M. Basic App
Myol 1996;6(2):70,115.
18 Masuda T Miyano H Sadoyama T EEG Clin Neurophysiol
108 ELECTROMYOGRAPHY
http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?-http://-/?- -
8/12/2019 Electromyography Medical Encyclopedia.pdf
12/12